thebench.bio #1 Ecosystem for the Bio-Tech Industry
All Quality Control questions

Question 17 of 20 Advanced Scenario

How would you handle an OOS result when the original worksheet contains a calculation error?

Short answer

Document it as an OOS investigation, not a quiet correction. Recalculate from the original raw data, confirm the error is real and fully explains the failure, have the correction reviewed and approved, and keep the original record. Then ask why the error was possible and whether other results are affected.

Why

The FDA's OOS guidance says the supervisor's first assessment should verify that the calculations converting raw data into a result are scientifically sound and correct, and check for unauthorized or unvalidated changes to automated calculations. A confirmed calculation error is one of the clearest laboratory errors, and the guidance says that when clear evidence of laboratory error exists, the results should be invalidated. Clear is the key word.

The correction must come from the original raw data, such as peak areas, weights and dilution factors, and must fully account for the OOS. If a dilution factor of 50 was typed as 5.0, recalculating should reproduce a passing value exactly from the same data. If the corrected result is still borderline or the error explains only part of the shift, the OOS is not resolved and the full investigation continues.

All original data must be kept with the explanation, initials and dates. The guidance also warns that laboratory error should be relatively rare, and frequent errors point to training, equipment or careless work. So the investigation asks why a manual calculation was possible.

A locked, validated template or system calculation is the CAPA, and other results calculated the same way need review, because the same error could have made a failing lot pass.

How to work it through

  1. Open the OOS record and preserve the original worksheet and raw data.
  2. Recalculate the result independently from raw values and identify the exact error.
  3. Confirm the corrected value fully explains the failure and is not borderline.
  4. Have the correction reviewed and approved by the supervisor and QA, with the original retained.
  5. Search other results calculated with the same worksheet or method for the same error.
  6. Implement a calculation control and record the CAPA.

An example

An assay reports 47.6 percent against 95.0 to 105.0 percent. The supervisor rechecks and finds the analyst entered a dilution factor of 2.5 instead of the 5.0 required by the method. Recalculating from the recorded peak areas and weights gives 95.2 percent.

That is in specification but close to the limit, so the team also reinjects the retained solution, which gives 95.4 percent, consistent with the corrected value. A search of 60 earlier results using the same template finds one more dilution error, which had produced a high but passing value. Both are corrected under the investigation and the template is locked.

Where the answer stops holding

A calculation error that only partly explains the result does not invalidate it. Correcting the arithmetic cannot fix errors in the raw data itself, such as a wrong weight recorded. Unlocked spreadsheets make it hard to prove which version of a formula was used, which weakens the conclusion.

A strong answer covers

  • Documented OOS investigation, not a silent fix
  • Recalculation from original raw data
  • Error must fully explain the failure
  • Original data retained with approval of the correction
  • Look for the same error in other results

Common mistakes

  • Correcting the worksheet and reporting the new value without opening an OOS investigation.
  • Accepting a correction that only partly explains the out-of-specification value.
  • Not checking whether the same error made other failing lots appear to pass.